Data Mining: 4. Algoritma Klasifikasi

Slides:



Advertisements
Presentasi serupa
Chapter 11 k- Fold Cross Validation
Advertisements

Modul-8 : Algoritma dan Struktur Data
2. Introduction to Algorithm and Programming
Chapter 9 ALGORITME Cluster dan WEKA
NoOUTLOKTEMPERATUREHUMIDITYWINDYPLAY 1SunnyHotHighFALSENo 2SunnyHotHighTRUENo 3CloudyHotHighFALSEYes 4RainyMildHighFALSEYes 5RainyCoolNormalFALSEYes 6RainyCoolNormalTRUEYes.
MULTILAYER PERCEPTRON
BLACK BOX TESTING.
Presented By : Group 2. A solution of an equation in two variables of the form. Ax + By = C and Ax + By + C = 0 A and B are not both zero, is an ordered.
Testing Implementasi Sistem Oleh :Rifiana Arief, SKom, MMSI
Data Mining: Klasifikasi dan Prediksi Naive Bayesian & Bayesian Network . April 13, 2017.
Pertemuan 05 Sebaran Peubah Acak Diskrit
Ruang Contoh dan Peluang Pertemuan 05
Pendugaan Parameter Proporsi dan Varians (Ragam) Pertemuan 14 Matakuliah: L0104 / Statistika Psikologi Tahun : 2008.
1 Pertemuan 10 Statistical Reasoning Matakuliah: T0264/Inteligensia Semu Tahun: Juli 2006 Versi: 2/1.
Masalah Transportasi II (Transportation Problem II)
Pertemuan XIV FUNGSI MAYOR Assosiation. What Is Association Mining? Association rule mining: –Finding frequent patterns, associations, correlations, or.
Pertemuan 07 Peluang Beberapa Sebaran Khusus Peubah Acak Kontinu
HAMPIRAN NUMERIK SOLUSI PERSAMAAN NIRLANJAR Pertemuan 3
NIPRL 1.4 Probabilitas Bersyarat Definisi Probabilitas Bersyarat(1/2) Probabilitas Bersyarat Probabilitas bersyarat kejadian A pada kejadian B adalah.
MULTIPLE REGRESSION ANALYSIS THE THREE VARIABLE MODEL: NOTATION AND ASSUMPTION 08/06/2015Ika Barokah S.
1 HAMPIRAN NUMERIK SOLUSI PERSAMAAN LANJAR Pertemuan 5 Matakuliah: K0342 / Metode Numerik I Tahun: 2006 TIK:Mahasiswa dapat meghitung nilai hampiran numerik.
1 Pertemuan 13 Algoritma Pergantian Page Matakuliah: T0316/sistem Operasi Tahun: 2005 Versi/Revisi: 5.
Data Mining: 7. Algoritma Estimasi
2-Pembelajaran Statistik 25 Agustus 2015 Data pelatihan dan pengujian Bias dan variansi Error rate & confidence interval Regresi Linear Praktikum: Data.
4- Classification: Logistic Regression 9 September 2015 Intro to Logistic Regression.
Ukuran Pemusatan dan Lokasi Pertemuan 03 Matakuliah: L0104 / Statistika Psikologi Tahun : 2008.
Expectation Maximization. Coin flipping experiment  Diberikan koin A dan B dengan nilai bias A dan B yang belum diketahui  Koin A akan memunculkan head.
Chapter 10 – The Design of Feedback Control Systems PID Compensation Networks.
Data Mining: 4. Algoritma Klasifikasi
Keuangan dan Akuntansi Proyek Modul 2: BASIC TOOLS CHRISTIONO UTOMO, Ph.D. Bidang Manajemen Proyek ITS 2011.
Smoothing. Basic Smoothing Models Moving average, weighted moving average, exponential smoothing Single and Double Smoothing First order exponential smoothing.
Naive Bayesian & Bayesian Network
PROBABILITY DISTRIBUTION
Klasifikasi Data Mining.
STATISTIKA CHATPER 4 (Perhitungan Dispersi (Sebaran))
Statistik TP A Pengujian Hipotesis Satu Populasi (Mean dan Proporsi)
Pert. 16. Menyimak lingkungan IS/IT saat ini
Algoritma C4.5. Algoritma C4.5 Object-Oriented Programming Introduction Algoritma C4.5 merupakan algoritma yang digunakan.
Data Mining Junta Zeniarja, M.Kom, M.CS
Decision Tree Classification.
Decision Tree Classification.
Bayes’ Theorem: Basics
Statistika Chapter 4 Probability.
Pengujian Hipotesis (I) Pertemuan 11
Matakuliah : I0014 / Biostatistika Tahun : 2005 Versi : V1 / R1
Clustering.
Presentasi Statistika Dasar
Data Mining: Klasifikasi Naive Bayesian & Bayesian Network
PENDUGAAN PARAMETER Pertemuan 8
the formula for the standard deviation:
Pohon Keputusan (Decision Trees)
Pendugaan Parameter (II) Pertemuan 10
Classification Supervised learning.
Pertemuan 09 Pengujian Hipotesis 2
Decision Tree.
Oleh : Rahmat Robi Waliyansyah, M.Kom.
Manajemen Proyek Perangkat Lunak (MPPL)
Master data Management
Pertemuan 4 CLASS DIAGRAM.
Self-Organizing Network Model (SOM) Pertemuan 10
MIK | FAKULTAS ILMU-ILMU KESEHATAN
Pertemuan 21 dan 22 Analisis Regresi dan Korelasi Sederhana
Simultaneous Linear Equations
Data Mining: Klasifikasi dan Prediksi Naive Bayesian & Bayesian Network . November 8, 2018.
Konsep Aplikasi Data Mining
CSG3G3 Kercerdasan Mesin dan Artifisial Pertemuan 10: Learning (Naive Bayes) Author-Tim Dosen KK ICM 21/11/2018.
KLASIFIKASI.
By Yulius Suprianto Macroeconomics | 02 Maret 2019 Chapter-5: The Standard of Living Over Time and A Cross Countries Source: http//
Data Mining Classification.
Universitas Gunadarma
Transcript presentasi:

Data Mining: 4. Algoritma Klasifikasi romi@romisatriawahono.net Data Mining: 4. Algoritma Klasifikasi Romi Satria Wahono romi@romisatriawahono.net http://romisatriawahono.net/dm WA/SMS: +6281586220090 http://romisatriawahono.net

Romi Satria Wahono SD Sompok Semarang (1987) SMPN 8 Semarang (1990) SMA Taruna Nusantara Magelang (1993) B.Eng, M.Eng and Ph.D in Software Engineering from Saitama University Japan (1994-2004) Universiti Teknikal Malaysia Melaka (2014) Research Interests: Software Engineering, Machine Learning Founder dan Koordinator IlmuKomputer.Com Peneliti LIPI (2004-2007) Founder dan CEO PT Brainmatics Cipta Informatika

Course Outline 1. Pengantar Data Mining 2. Proses Data Mining 3. Persiapan Data 4. Algoritma Klasifikasi 5. Algoritma Klastering 6. Algoritma Asosiasi 7. Algoritma Estimasi

4. Algoritma Klasifikasi 4.1 Decision Tree Induction 4.2 Bayesian Classification 4.3 Neural Network 4.4 Model Evaluation and Selection 4.5 Techniques to Improve Classification Accuracy: Ensemble Methods

4.1 Decision Tree

Algorithm for Decision Tree Induction Basic algorithm (a greedy algorithm) Tree is constructed in a top-down recursive divide-and-conquer manner At start, all the training examples are at the root Attributes are categorical (if continuous-valued, they are discretized in advance) Examples are partitioned recursively based on selected attributes Test attributes are selected on the basis of a heuristic or statistical measure (e.g., information gain, gain ratio, gini index) Conditions for stopping partitioning All samples for a given node belong to the same class There are no remaining attributes for further partitioning – majority voting is employed for classifying the leaf There are no samples left

Brief Review of Entropy   m = 2

Attribute Selection Measure: Information Gain (ID3/C4.5) Select the attribute with the highest information gain Let pi be the probability that an arbitrary tuple in D belongs to class Ci, estimated by | Ci, D|/|D| Expected information (entropy) needed to classify a tuple in D: Information needed (after using A to split D into v partitions) to classify D: Information gained by branching on attribute A

Attribute Selection: Information Gain Class P: buys_computer = “yes” Class N: buys_computer = “no” means “age <=30” has 5 out of 14 samples, with 2 yes’es and 3 no’s. Hence Similarly,

Computing Information-Gain for Continuous-Valued Attributes Let attribute A be a continuous-valued attribute Must determine the best split point for A Sort the value A in increasing order Typically, the midpoint between each pair of adjacent values is considered as a possible split point (ai+ai+1)/2 is the midpoint between the values of ai and ai+1 The point with the minimum expected information requirement for A is selected as the split-point for A Split: D1 is the set of tuples in D satisfying A ≤ split-point, and D2 is the set of tuples in D satisfying A > split-point

Tahapan Algoritma Decision Tree Siapkan data training Pilih atribut sebagai akar Buat cabang untuk tiap-tiap nilai Ulangi proses untuk setiap cabang sampai semua kasus pada cabang memiliki kelas yg sama

1. Siapkan data training

2. Pilih atribut sebagai akar Untuk memilih atribut akar, didasarkan pada nilai Gain tertinggi dari atribut-atribut yang ada. Untuk mendapatkan nilai Gain, harus ditentukan terlebih dahulu nilai Entropy Rumus Entropy: S = Himpunan Kasus n = Jumlah Partisi S pi = Proporsi dari Si terhadap S Rumus Gain: S = Himpunan Kasus A = Atribut n = Jumlah Partisi Atribut A | Si | = Jumlah Kasus pada partisi ke-i | S | = Jumlah Kasus dalam S

Perhitungan Entropy dan Gain Akar

Penghitungan Entropy Akar Entropy Total Entropy (Outlook) Entropy (Temperature) Entropy (Humidity) Entropy (Windy)

Penghitungan Entropy Akar NODE ATRIBUT JML KASUS (S) YA (Si) TIDAK (Si) ENTROPY GAIN 1 TOTAL   14 10 4 0,86312 OUTLOOK CLOUDY RAINY 5 0,72193 SUNNY 2 3 0,97095 TEMPERATURE COOL HOT MILD 6 0,91830 HUMADITY HIGH 7 0,98523 NORMAL WINDY FALSE 8 0,81128 TRUE

Penghitungan Gain Akar

Penghitungan Gain Akar NODE ATRIBUT JML KASUS (S) YA (Si) TIDAK (Si) ENTROPY GAIN 1 TOTAL   14 10 4 0,86312 OUTLOOK 0,25852 CLOUDY RAINY 5 0,72193 SUNNY 2 3 0,97095 TEMPERATURE 0,18385 COOL HOT MILD 6 0,91830 HUMADITY 0,37051 HIGH 7 0,98523 NORMAL WINDY 0,00598 FALSE 8 0,81128 TRUE

Gain Tertinggi Sebagai Akar Dari hasil pada Node 1, dapat diketahui bahwa atribut dengan Gain tertinggi adalah HUMIDITY yaitu sebesar 0.37051 Dengan demikian HUMIDITY dapat menjadi node akar Ada 2 nilai atribut dari HUMIDITY yaitu HIGH dan NORMAL. Dari kedua nilai atribut tersebut, nilai atribut NORMAL sudah mengklasifikasikan kasus menjadi 1 yaitu keputusan-nya Yes, sehingga tidak perlu dilakukan perhitungan lebih lanjut Tetapi untuk nilai atribut HIGH masih perlu dilakukan perhitungan lagi 1. HUMIDITY 1.1 ????? Yes High Normal

2. Buat cabang untuk tiap-tiap nilai Untuk memudahkan, dataset di filter dengan mengambil data yang memiliki kelembaban HUMADITY=HIGH untuk membuat table Node 1.1 OUTLOOK TEMPERATURE HUMIDITY WINDY PLAY Sunny Hot High FALSE No TRUE Cloudy Yes Rainy Mild

Perhitungan Entropi Dan Gain Cabang NODE ATRIBUT JML KASUS (S) YA (Si) TIDAK (Si) ENTROPY GAIN 1.1 HUMADITY   7 3 4 0,98523 OUTLOOK 0,69951 CLOUDY 2 RAINY 1 SUNNY TEMPERATURE 0,02024 COOL HOT 0,91830 MILD WINDY FALSE TRUE

Gain Tertinggi Sebagai Node 1.1 Dari hasil pada Tabel Node 1.1, dapat diketahui bahwa atribut dengan Gain tertinggi adalah OUTLOOK yaitu sebesar 0.69951 Dengan demikian OUTLOOK dapat menjadi node kedua Artibut CLOUDY = YES dan SUNNY= NO sudah mengklasifikasikan kasus menjadi 1 keputusan, sehingga tidak perlu dilakukan perhitungan lebih lanjut Tetapi untuk nilai atribut RAINY masih perlu dilakukan perhitungan lagi 1. HUMIDITY 1.1 OUTLOOK Yes High Normal No 1.1.2 ????? Cloudy Rainy Sunny

3. Ulangi proses untuk setiap cabang sampai semua kasus pada cabang memiliki kelas yg sama OUTLOOK TEMPERATURE HUMIDITY WINDY PLAY Rainy Mild High FALSE Yes TRUE No NODE ATRIBUT JML KASUS (S) YA (Si) TIDAK (Si) ENTROPY GAIN 1.2 HUMADITY HIGH & OUTLOOK RAINY   2 1 TEMPERATURE COOL HOT MILD WINDY FALSE TRUE

Gain Tertinggi Sebagai Node 1.1.2 1. HUMIDITY 1.1 OUTLOOK Yes High Normal No 1.1.2 WINDY Cloudy Rainy Sunny False True Dari tabel, Gain Tertinggi adalah WINDY dan menjadi node cabang dari atribut RAINY Karena semua kasus sudah masuk dalam kelas Jadi, pohon keputusan pada Gambar merupakan pohon keputusan terakhir yang terbentuk

Decision Tree Induction: An Example age income student credit_rating buys_computer <=30 high no fair excellent 31…40 yes >40 medium low Training data set: Buys_computer

Gain Ratio for Attribute Selection (C4.5) Information gain measure is biased towards attributes with a large number of values C4.5 (a successor of ID3) uses gain ratio to overcome the problem (normalization to information gain) GainRatio(A) = Gain(A)/SplitInfo(A) Ex. gain_ratio(income) = 0.029/1.557 = 0.019 The attribute with the maximum gain ratio is selected as the splitting attribute

Gini Index (CART) If a data set D contains examples from n classes, gini index, gini(D) is defined as where pj is the relative frequency of class j in D If a data set D is split on A into two subsets D1 and D2, the gini index gini(D) is defined as Reduction in Impurity: The attribute provides the smallest ginisplit(D) (or the largest reduction in impurity) is chosen to split the node (need to enumerate all the possible splitting points for each attribute)

Computation of Gini Index Ex. D has 9 tuples in buys_computer = “yes” and 5 in “no” Suppose the attribute income partitions D into 10 in D1: {low, medium} and 4 in D2 Gini{low,high} is 0.458; Gini{medium,high} is 0.450. Thus, split on the {low,medium} (and {high}) since it has the lowest Gini index All attributes are assumed continuous-valued May need other tools, e.g., clustering, to get the possible split values Can be modified for categorical attributes

Comparing Attribute Selection Measures The three measures, in general, return good results but Information gain: biased towards multivalued attributes Gain ratio: tends to prefer unbalanced splits in which one partition is much smaller than the others Gini index: biased to multivalued attributes has difficulty when # of classes is large tends to favor tests that result in equal-sized partitions and purity in both partitions

Other Attribute Selection Measures CHAID: a popular decision tree algorithm, measure based on χ2 test for independence C-SEP: performs better than info. gain and gini index in certain cases G-statistic: has a close approximation to χ2 distribution MDL (Minimal Description Length) principle (i.e., the simplest solution is preferred): The best tree as the one that requires the fewest # of bits to both (1) encode the tree, and (2) encode the exceptions to the tree Multivariate splits (partition based on multiple variable combinations) CART: finds multivariate splits based on a linear comb. of attrs. Which attribute selection measure is the best? Most give good results, none is significantly superior than others

Overfitting and Tree Pruning Overfitting: An induced tree may overfit the training data Too many branches, some may reflect anomalies due to noise or outliers Poor accuracy for unseen samples Two approaches to avoid overfitting Prepruning: Halt tree construction early ̵ do not split a node if this would result in the goodness measure falling below a threshold Difficult to choose an appropriate threshold Postpruning: Remove branches from a “fully grown” tree -get a sequence of progressively pruned trees Use a set of data different from the training data to decide which is the “best pruned tree”

Pruning

Why is decision tree induction popular? Relatively faster learning speed (than other classification methods) Convertible to simple and easy to understand classification rules Can use SQL queries for accessing databases Comparable classification accuracy with other methods

Latihan Lakukan eksperimen mengikuti buku Matthew North (Data Mining for the Masses) Chapter Ten (Decision Tree) Analisis jenis decision tree apa saja yang digunakan dan mengapa perlu dilakukan pada dataset tersebut

4.2 Bayesian Classification

Bayesian Classification: Why? A statistical classifier: performs probabilistic prediction, i.e., predicts class membership probabilities Foundation: Based on Bayes’ Theorem. Performance: A simple Bayesian classifier, naïve Bayesian classifier, has comparable performance with decision tree and selected neural network classifiers Incremental: Each training example can incrementally increase/decrease the probability that a hypothesis is correct — prior knowledge can be combined with observed data Standard: Even when Bayesian methods are computationally intractable, they can provide a standard of optimal decision making against which other methods can be measured

Bayes’ Theorem: Basics Total probability Theorem: Bayes’ Theorem: Let X be a data sample (“evidence”): class label is unknown Let H be a hypothesis that X belongs to class C Classification is to determine P(H|X), (i.e., posteriori probability): the probability that the hypothesis holds given the observed data sample X P(H) (prior probability): the initial probability E.g., X will buy computer, regardless of age, income, … P(X): probability that sample data is observed P(X|H) (likelihood): the probability of observing the sample X, given that the hypothesis holds E.g., Given that X will buy computer, the prob. that X is 31..40, medium income

Prediction Based on Bayes’ Theorem Given training data X, posteriori probability of a hypothesis H, P(H|X), follows the Bayes’ theorem Informally, this can be viewed as posteriori = likelihood x prior/evidence Predicts X belongs to Ci iff the probability P(Ci|X) is the highest among all the P(Ck|X) for all the k classes Practical difficulty: It requires initial knowledge of many probabilities, involving significant computational cost

Classification is to Derive the Maximum Posteriori Let D be a training set of tuples and their associated class labels, and each tuple is represented by an n-D attribute vector X = (x1, x2, …, xn) Suppose there are m classes C1, C2, …, Cm. Classification is to derive the maximum posteriori, i.e., the maximal P(Ci|X) This can be derived from Bayes’ theorem Since P(X) is constant for all classes, only needs to be maximized

Naïve Bayes Classifier A simplified assumption: attributes are conditionally independent (i.e., no dependence relation between attributes): This greatly reduces the computation cost: Only counts the class distribution If Ak is categorical, P(xk|Ci) is the # of tuples in Ci having value xk for Ak divided by |Ci, D| (# of tuples of Ci in D) If Ak is continous-valued, P(xk|Ci) is usually computed based on Gaussian distribution with a mean μ and standard deviation σ and P(xk|Ci) is

Naïve Bayes Classifier: Training Dataset age income student credit_rating buys_computer <=30 high no fair excellent 31…40 yes >40 medium low No Yes Class: C1:buys_computer = ‘yes’ C2:buys_computer = ‘no’ Data to be classified: X = (age <=30, income = medium, student = yes, credit_rating = fair) X  buy computer?

Naïve Bayes Classifier: An Example age income student credit_rating buys_computer <=30 high no fair excellent 31…40 yes >40 medium low P(Ci): P(buys_computer = “yes”) = 9/14 = 0.643 P(buys_computer = “no”) = 5/14= 0.357 Compute P(X|Ci) for each class P(age = “<=30” | buys_computer = “yes”) = 2/9 = 0.222 P(age = “<= 30” | buys_computer = “no”) = 3/5 = 0.6 P(income = “medium” | buys_computer = “yes”) = 4/9 = 0.444 P(income = “medium” | buys_computer = “no”) = 2/5 = 0.4 P(student = “yes” | buys_computer = “yes) = 6/9 = 0.667 P(student = “yes” | buys_computer = “no”) = 1/5 = 0.2 P(credit_rating = “fair” | buys_computer = “yes”) = 6/9 = 0.667 P(credit_rating = “fair” | buys_computer = “no”) = 2/5 = 0.4 X = (age <= 30 , income = medium, student = yes, credit_rating = fair) P(X|Ci) : P(X|buys_computer = “yes”) = 0.222 x 0.444 x 0.667 x 0.667 = 0.044 P(X|buys_computer = “no”) = 0.6 x 0.4 x 0.2 x 0.4 = 0.019 P(X|Ci)*P(Ci) : P(X|buys_computer = “yes”) * P(buys_computer = “yes”) = 0.028 P(X|buys_computer = “no”) * P(buys_computer = “no”) = 0.007 Therefore, X belongs to class (“buys_computer = yes”)

Tahapan Algoritma Naïve Bayes Baca Data Training Hitung jumlah class Hitung jumlah kasus yang sama dengan class yang sama Kalikan semua nilai hasil sesuai dengan data X yang dicari class-nya

1. Baca Data Training

Teorema Bayes X  Data dengan class yang belum diketahui H  Hipotesis data X yang merupakan suatu class yang lebih spesifik P (H|X)  Probabilitas hipotesis H berdasarkan kondisi X (posteriori probability) P (H)  Probabilitas hipotesis H (prior probability) P (X|H)  Probabilitas X berdasarkan kondisi pada hipotesis H P (X)  Probabilitas X

2. Hitung jumlah class/label Terdapat 2 class dari data training tersebut, yaitu: C1 (Class 1) Play = yes  9 record C2 (Class 2) Play = no  5 record Total = 14 record Maka: P (C1) = 9/14 = 0.642857143 P (C2) = 5/14 = 0.357142857 Pertanyaan: Data X = (outlook=rainy, temperature=cool, humidity=high, windy=true) Main golf atau tidak?

3. Hitung jumlah kasus yang sama dengan class yang sama Untuk P(Ci) yaitu P(C1) dan P(C2) sudah diketahui hasilnya di langkah sebelumnya. Selanjutnya Hitung P(X|Ci) untuk i = 1 dan 2 P(outlook=“sunny”|play=“yes”)=2/9=0.222222222 P(outlook=“sunny”|play=“no”)=3/5=0.6 P(outlook=“overcast”|play=“yes”)=4/9=0.444444444 P(outlook=“overcast”|play=“no”)=0/5=0 P(outlook=“rainy”|play=“yes”)=3/9=0.333333333 P(outlook=“rainy”|play=“no”)=2/5=0.4

3. Hitung jumlah kasus yang sama dengan class yang sama romi@romisatriawahono.net Object-Oriented Programming 3. Hitung jumlah kasus yang sama dengan class yang sama Jika semua atribut dihitung, maka didapat hasil akhirnya seperti berikut ini: Atribute Parameter No Yes Outlook value=sunny 0.6 0.2222222222222222 value=cloudy 0.0 0.4444444444444444 value=rainy 0.4 0.3333333333333333 Temperature value=hot value=mild value=cool 0.2 Humidity value=high 0.8 value=normal 0.6666666666666666 Windy value=false value=true http://romisatriawahono.net

4. Kalikan semua nilai hasil sesuai dengan data X yang dicari class-nya Pertanyaan: Data X = (outlook=rainy, temperature=cool, humidity=high, windy=true) Main Golf atau tidak? Kalikan semua nilai hasil dari data X P(X|play=“yes”) = 0.333333333* 0.333333333* 0.333333333*0.333333333 = 0.012345679 P(X|play=“no”) = 0.4*0.2*0.8*0.6=0.0384 P(X|play=“yes”)*P(C1) = 0.012345679*0.642857143 = 0.007936508 P(X|play=“no”)*P(C2) = 0.0384*0.357142857 = 0.013714286 Nilai “no” lebih besar dari nilai “yes” maka class dari data X tersebut adalah “No”

Avoiding the Zero-Probability Problem Naïve Bayesian prediction requires each conditional prob. be non-zero. Otherwise, the predicted prob. will be zero Ex. Suppose a dataset with 1000 tuples, income=low (0), income= medium (990), and income = high (10) Use Laplacian correction (or Laplacian estimator) Adding 1 to each case Prob(income = low) = 1/1003 Prob(income = medium) = 991/1003 Prob(income = high) = 11/1003 The “corrected” prob. estimates are close to their “uncorrected” counterparts

Naïve Bayes Classifier: Comments Advantages Easy to implement Good results obtained in most of the cases Disadvantages Assumption: class conditional independence, therefore loss of accuracy Practically, dependencies exist among variables, e.g.: Hospitals Patients Profile: age, family history, etc. Symptoms: fever, cough etc., Disease: lung cancer, diabetes, etc. Dependencies among these cannot be modeled by Naïve Bayes Classifier How to deal with these dependencies? Bayesian Belief Networks

4.3 Neural Network

Neural Network Neural Network adalah suatu model yang dibuat untuk meniru fungsi belajar yang dimiliki otak manusia atau jaringan dari sekelompok unit pemroses kecil yang dimodelkan berdasarkan jaringan saraf manusia

Neural Network Model Perceptron adalah model jaringan yang terdiri dari beberapa unit masukan (ditambah dengan sebuah bias), dan memiliki sebuah unit keluaran Fungsi aktivasi bukan hanya merupakan fungsi biner (0,1) melainkan bipolar (1,0,-1) Untuk suatu harga threshold ѳ yang ditentukan: 1 Jika net > ѳ F (net) = 0 Jika – ѳ ≤ net ≤ ѳ -1 Jika net < - ѳ

Fungsi Aktivasi Macam fungsi aktivasi yang dipakai untuk mengaktifkan net diberbagai jenis neural network: Aktivasi linear, Rumus: y = sign(v) = v Aktivasi step, Rumus: Aktivasi sigmoid biner, Rumus: Aktivasi sigmoid bipolar, Rumus:

Tahapan Algoritma Perceptron Inisialisasi semua bobot dan bias (umumnya wi = b = 0) Selama ada element vektor masukan yang respon unit keluarannya tidak sama dengan target, lakukan: 2.1 Set aktivasi unit masukan xi = Si (i = 1,...,n) 2.2 Hitung respon unit keluaran: net = + b 1 Jika net > ѳ F (net) = 0 Jika – ѳ ≤ net ≤ ѳ -1 Jika net < - ѳ 2.3 Perbaiki bobot pola yang mengadung kesalahan menurut persamaan: Wi (baru) = wi (lama) + ∆w (i = 1,...,n) dengan ∆w = α t xi B (baru) = b(lama) + ∆ b dengan ∆b = α t Dimana: α = Laju pembelajaran (Learning rate) yang ditentukan ѳ = Threshold yang ditentukan t = Target 2.4 Ulangi iterasi sampai perubahan bobot (∆wn = 0) tidak ada

Studi Kasus Diketahui sebuah dataset kelulusan berdasarkan IPK untuk program S1: Jika ada mahasiswa IPK 2.85 dan masih semester 1, maka masuk ke kedalam manakah status tersebut ? Status IPK Semester Lulus 2.9 1 Tidak Lulus 2.8 3 2.3 5 Tidak lulus 2.7 6

1: Inisialisasi Bobot Inisialisasi Bobot dan bias awal: b = 0 dan bias = 1 t X1 X2 1 2,9 -1 2.8 3 2.3 5 2,7 6

2.1: Set aktivasi unit masukan Treshold (batasan), θ = 0 , yang artinya : 1 Jika net > 0 F (net) = 0 Jika net = 0 -1 Jika net < 0

2.2 - 2.3 Hitung Respon dan Perbaiki Bobot Hitung Response Keluaran iterasi 1 Perbaiki bobot pola yang mengandung kesalahan MASUKAN TARGET   y= PERUBAHAN BOBOT BOBOT BARU X1 X2 1 t NET f(NET) ∆W1 ∆W2 ∆b W1 W2 b INISIALISASI 2,9 7 2,8 3 -1 8,12 -2,8 -3 0,1 4 2,3 5 0,23 -2,3 -5 -2,2 2,7 6 -5,94

2.4 Ulangi iterasi sampai perubahan bobot (∆wn = 0) tidak ada (Iterasi 2) Hitung Response Keluaran iterasi 2 Perbaiki bobot pola yang mengandung kesalahan MASUKAN TARGET   y= PERUBAHAN BOBOT BOBOT BARU X1 X2 1 t NET f(NET) ∆W1 ∆W2 ∆b W1 W2 b INISIALISASI -2,2 -1 2,9 -8,38 0,7 2,8 3 1,96 -2,8 -3 -2,1 2,3 5 -20,83 2,7 6 -24,67

2.4 Ulangi iterasi sampai perubahan bobot (∆wn = 0) tidak ada (Iterasi 3) Hitung Response Keluaran iterasi 3 Perbaiki bobot pola yang mengandung kesalahan Untuk data IPK memiliki pola 0.8 x - 2 y = 0 dapat dihitung prediksinya menggunakan bobot yang terakhir didapat: V = X1*W1 + X2*W2 = 0,8 * 2,85 - 2*1 = 2,28 -2 = 0,28 Y = sign(V) = sign(0,28) = 1 (Lulus) MASUKAN TARGET   y= PERUBAHAN BOBOT BOBOT BARU X1 X2 1 t NET f(NET) ∆W1 ∆W2 ∆b W1 W2 b INISIALISASI -2,1 -3 -1 2,9 -10,09 0,8 -2 2,8 3 -3,76 2,3 5 -8,16 2,7 6 -9,84

4.4 Model Evaluation and Selection

Model Evaluation and Selection Evaluation metrics: How can we measure accuracy? Other metrics to consider? Use validation test set of class-labeled tuples instead of training set when assessing accuracy Methods for estimating a classifier’s accuracy: Holdout method, random subsampling Cross-validation Bootstrap Comparing classifiers: Confidence intervals Cost-benefit analysis and ROC Curves

Classifier Evaluation Metrics: Confusion Matrix Actual class\Predicted class C1 ¬ C1 True Positives (TP) False Negatives (FN) False Positives (FP) True Negatives (TN) Actual class\Predicted class buy_computer = yes buy_computer = no Total buy_computer = yes 6954 46 7000 412 2588 3000 7366 2634 10000 Given m classes, an entry, CMi,j in a confusion matrix indicates # of tuples in class i that were labeled by the classifier as class j May have extra rows/columns to provide totals

Classifier Evaluation Metrics: Accuracy, Error Rate, Sensitivity and Specificity Class Imbalance Problem: One class may be rare, e.g. fraud, or HIV-positive Significant majority of the negative class and minority of the positive class Sensitivity: True Positive recognition rate Sensitivity = TP/P Specificity: True Negative recognition rate Specificity = TN/N A\P C ¬C TP FN P FP TN N P’ N’ All Classifier Accuracy or recognition rate: percentage of test set tuples that are correctly classified Accuracy = (TP + TN)/All Error rate: 1 – accuracy, or Error rate = (FP + FN)/All

Classifier Evaluation Metrics: Precision and Recall, and F-measures Precision: exactness – what % of tuples that the classifier labeled as positive are actually positive Recall: completeness – what % of positive tuples did the classifier label as positive? Perfect score is 1.0 Inverse relationship between precision & recall F measure (F1 or F-score): harmonic mean of precision and recall, Fß: weighted measure of precision and recall assigns ß times as much weight to recall as to precision

Classifier Evaluation Metrics: Example Actual Class\Predicted class cancer = yes cancer = no Total Recognition(%) 90 210 300 30.00 (sensitivity 140 9560 9700 98.56 (specificity) 230 9770 10000 96.40 (accuracy) Precision = 90/230 = 39.13% Recall = 90/300 = 30.00%

Evaluating Classifier Accuracy: Holdout & Cross-Validation Methods Holdout method Given data is randomly partitioned into two independent sets Training set (e.g., 2/3) for model construction Test set (e.g., 1/3) for accuracy estimation Random sampling: a variation of holdout Repeat holdout k times, accuracy = avg. of the accuracies obtained Cross-validation (k-fold, where k = 10 is most popular) Randomly partition the data into k mutually exclusive subsets, each approximately equal size At i-th iteration, use Di as test set and others as training set Leave-one-out: k folds where k = # of tuples, for small sized data *Stratified cross-validation*: folds are stratified so that class dist. in each fold is approx. the same as that in the initial data

Evaluating Classifier Accuracy: Bootstrap Works well with small data sets Samples the given training tuples uniformly with replacement, i.e., each time a tuple is selected, it is equally likely to be selected again and re- added to the training set Several bootstrap methods, and a common one is .632 boostrap A data set with d tuples is sampled d times, with replacement, resulting in a training set of d samples The data tuples that did not make it into the training set end up forming the test set. About 63.2% of the original data end up in the bootstrap, and the remaining 36.8% form the test set (since (1 – 1/d)d ≈ e-1 = 0.368) Repeat the sampling procedure k times, overall accuracy of the model:

Estimating Confidence Intervals: Classifier Models M1 vs. M2 Suppose we have two classifiers, M1 and M2, which one is better? Use 10-fold cross-validation to obtain and These mean error rates are just estimates of error on the true population of future data cases What if the difference between the two error rates is just attributed to chance? Use a test of statistical significance Obtain confidence limits for our error estimates

Estimating Confidence Intervals: Null Hypothesis Perform 10-fold cross-validation Assume samples follow a t distribution with k–1 degrees of freedom (here, k=10) Use t-test (or Student’s t-test) Null Hypothesis: M1 & M2 are the same If we can reject null hypothesis, then we conclude that the difference between M1 & M2 is statistically significant Chose model with lower error rate

Estimating Confidence Intervals: t-test If only 1 test set available: pairwise comparison For ith round of 10-fold cross-validation, the same cross partitioning is used to obtain err(M1)i and err(M2)i Average over 10 rounds to get t-test computes t-statistic with k-1 degrees of freedom: If two test sets available: use non-paired t-test where where where k1 & k2 are # of cross-validation samples used for M1 & M2, resp.

Estimating Confidence Intervals: Table for t-distribution Symmetric Significance level, e.g., sig = 0.05 or 5% means M1 & M2 are significantly different for 95% of population Confidence limit, z = sig/2

Estimating Confidence Intervals: Statistical Significance Are M1 & M2 significantly different? Compute t. Select significance level (e.g. sig = 5%) Consult table for t-distribution: Find t value corresponding to k-1 degrees of freedom (here, 9) t-distribution is symmetric: typically upper % points of distribution shown → look up value for confidence limit z=sig/2 (here, 0.025) If t > z or t < -z, then t value lies in rejection region: Reject null hypothesis that mean error rates of M1 & M2 are same Conclude: statistically significant difference between M1 & M2 Otherwise, conclude that any difference is chance

Model Selection: ROC Curves ROC (Receiver Operating Characteristics) curves: for visual comparison of classification models Originated from signal detection theory Shows the trade-off between the true positive rate and the false positive rate The area under the ROC curve is a measure of the accuracy of the model Rank the test tuples in decreasing order: the one that is most likely to belong to the positive class appears at the top of the list The closer to the diagonal line (i.e., the closer the area is to 0.5), the less accurate is the model Vertical axis represents the true positive rate Horizontal axis rep. the false positive rate The plot also shows a diagonal line A model with perfect accuracy will have an area of 1.0

Issues Affecting Model Selection Accuracy classifier accuracy: predicting class label Speed time to construct the model (training time) time to use the model (classification/prediction time) Robustness: handling noise and missing values Scalability: efficiency in disk-resident databases Interpretability understanding and insight provided by the model Other measures, e.g., goodness of rules, such as decision tree size or compactness of classification rules

4.5 Techniques to Improve Classification Accuracy: Ensemble Methods

Ensemble Methods: Increasing the Accuracy Use a combination of models to increase accuracy Combine a series of k learned models, M1, M2, …, Mk, with the aim of creating an improved model M* Popular ensemble methods Bagging: averaging the prediction over a collection of classifiers Boosting: weighted vote with a collection of classifiers Ensemble: combining a set of heterogeneous classifiers

Bagging: Boostrap Aggregation Analogy: Diagnosis based on multiple doctors’ majority vote Training Given a set D of d tuples, at each iteration i, a training set Di of d tuples is sampled with replacement from D (i.e., bootstrap) A classifier model Mi is learned for each training set Di Classification: classify an unknown sample X Each classifier Mi returns its class prediction The bagged classifier M* counts the votes and assigns the class with the most votes to X Prediction: can be applied to the prediction of continuous values by taking the average value of each prediction for a given test tuple Accuracy Often significantly better than a single classifier derived from D For noise data: not considerably worse, more robust Proved improved accuracy in prediction

Boosting Analogy: Consult several doctors, based on a combination of weighted diagnoses—weight assigned based on the previous diagnosis accuracy How boosting works? Weights are assigned to each training tuple A series of k classifiers is iteratively learned After a classifier Mi is learned, the weights are updated to allow the subsequent classifier, Mi+1, to pay more attention to the training tuples that were misclassified by Mi The final M* combines the votes of each individual classifier, where the weight of each classifier's vote is a function of its accuracy Boosting algorithm can be extended for numeric prediction Comparing with bagging: Boosting tends to have greater accuracy, but it also risks overfitting the model to misclassified data

Adaboost (Freund and Schapire, 1997) Given a set of d class-labeled tuples, (X1, y1), …, (Xd, yd) Initially, all the weights of tuples are set the same (1/d) Generate k classifiers in k rounds. At round i, Tuples from D are sampled (with replacement) to form a training set Di of the same size Each tuple’s chance of being selected is based on its weight A classification model Mi is derived from Di Its error rate is calculated using Di as a test set If a tuple is misclassified, its weight is increased, o.w. it is decreased Error rate: err(Xj) is the misclassification error of tuple Xj. Classifier Mi error rate is the sum of the weights of the misclassified tuples: The weight of classifier Mi’s vote is

Random Forest (Breiman 2001) Each classifier in the ensemble is a decision tree classifier and is generated using a random selection of attributes at each node to determine the split During classification, each tree votes and the most popular class is returned Two Methods to construct Random Forest: Forest-RI (random input selection): Randomly select, at each node, F attributes as candidates for the split at the node. The CART methodology is used to grow the trees to maximum size Forest-RC (random linear combinations): Creates new attributes (or features) that are a linear combination of the existing attributes (reduces the correlation between individual classifiers) Comparable in accuracy to Adaboost, but more robust to errors and outliers Insensitive to the number of attributes selected for consideration at each split, and faster than bagging or boosting

Classification of Class-Imbalanced Data Sets Class-imbalance problem: Rare positive example but numerous negative ones, e.g., medical diagnosis, fraud, oil-spill, fault, etc. Traditional methods assume a balanced distribution of classes and equal error costs: not suitable for class-imbalanced data Typical methods for imbalance data in 2-class classification: Oversampling: re-sampling of data from positive class Under-sampling: randomly eliminate tuples from negative class Threshold-moving: moves the decision threshold, t, so that the rare class tuples are easier to classify, and hence, less chance of costly false negative errors Ensemble techniques: Ensemble multiple classifiers introduced above Still difficult for class imbalance problem on multiclass tasks

Rangkuman Classification is a form of data analysis that extracts models describing important data classes Effective and scalable methods have been developed for decision tree induction, Naive Bayesian classification, rule-based classification, and many other classification methods Evaluation metrics include: accuracy, sensitivity, specificity, precision, recall, F measure, and Fß measure Stratified k-fold cross-validation is recommended for accuracy estimation. Bagging and boosting can be used to increase overall accuracy by learning and combining a series of individual models

Rangkuman Significance tests and ROC curves are useful for model selection. There have been numerous comparisons of the different classification methods; the matter remains a research topic No single method has been found to be superior over all others for all data sets Issues such as accuracy, training time, robustness, scalability, and interpretability must be considered and can involve trade-offs, further complicating the quest for an overall superior method

romi@romisatriawahono.net Object-Oriented Programming Referensi Jiawei Han and Micheline Kamber, Data Mining: Concepts and Techniques Third Edition, Elsevier, 2012 Ian H. Witten, Frank Eibe, Mark A. Hall, Data mining: Practical Machine Learning Tools and Techniques 3rd Edition, Elsevier, 2011 Markus Hofmann and Ralf Klinkenberg, RapidMiner: Data Mining Use Cases and Business Analytics Applications, CRC Press Taylor & Francis Group, 2014 Daniel T. Larose, Discovering Knowledge in Data: an Introduction to Data Mining, John Wiley & Sons, 2005 Ethem Alpaydin, Introduction to Machine Learning, 3rd ed., MIT Press, 2014 Florin Gorunescu, Data Mining: Concepts, Models and Techniques, Springer, 2011 Oded Maimon and Lior Rokach, Data Mining and Knowledge Discovery Handbook Second Edition, Springer, 2010 Warren Liao and Evangelos Triantaphyllou (eds.), Recent Advances in Data Mining of Enterprise Data: Algorithms and Applications, World Scientific, 2007 http://romisatriawahono.net